Re: remerge of msvc re-raise exception patch from Alberto

[Alexandre Julliard]

Greg Turner  writes:
> Another unmerged patch I'd like to see go in.  cppexcept.c has changed a 
> bit since this was submitted so I've tried to merge in the changes (at 
> a glance, this seems to have worked).  Let me know if there's anything 
> that needs fixing before it can go in.

This one should no longer be necessary. If you still have exception
troubles with the current CVS please send me a trace.

remerge of msvc re-raise exception patch from Alberto

[Greg Turner]

On Wednesday 13 November 2002 01:07 pm, Alexandre Julliard wrote:
> Greg Turner  writes:
> > Another unmerged patch I'd like to see go in.  cppexcept.c has
> > changed a bit since this was submitted so I've tried to merge in
> > the changes (at a glance, this seems to have worked).  Let me know
> > if there's anything that needs fixing before it can go in.
>
> This one should no longer be necessary. If you still have exception
> troubles with the current CVS please send me a trace.

hmmm... I've only run this through the debugger once, and even then I 
just observed the ever-increasing stack size and deduced that this was 
not handled... but I guess that explains why, IIRC, I was seeing bad 
pointer references instead of NULL one's as before.

Should a +seh trace be sufficient for a meaningful investigation?  If 
so, I can whip up a quite tidy (iow, small) one, with some annotations 
describing what I'm doing with the app, and post it to the list.

Never mind... before I bother wasting everyone's time (or even just 
yours), I'll take a closer look at this myself; maybe there's something 
obvious (or perhaps even something mildly esoteric) that I can fix, or 
at least diagnose, myself.  If not, at least some time in the debugger 
will hopefully clarify what's going on just before TSHTF, which might 
make a trace eaiser for whomever to interpret.

thanks,

remerge of msvc re-raise exception patch from Alberto

[Alexandre Julliard]

Greg Turner  writes:
> Should a +seh trace be sufficient for a meaningful investigation?  If 
> so, I can whip up a quite tidy (iow, small) one, with some annotations 
> describing what I'm doing with the app, and post it to the list.

+seh is usually enough, though +relay,+seh is sometimes necessary.
Also if you have a binary that you can send me that helps a lot.